For sound generation/effects you should read some book on DSP...It's really exciting having your own sound engine based on synthesis...you have full control on the sound. It can be dinamically adapted to the situation of the game ecc.
If you want you can use stk: http://ccrma.stanford.edu/software/stk/
have a read at this classic book (free!): http://www.dspguide.com/pdfbook.htm
anyway making a game based on sound files (like many games do) is much more simpler...so think about it.
Generating your own sound effects will fall well short of where you want your sound effects to be. Most games will use pre-recorded sounds and then post mix them using Sound Forge or something similar. Games that generate their sounds on the fly (such as driving games, flight sim games, etc.) often sound very fake. Test Drive Unlimited generates all their engine sounds and as a result they sound a bit unrealistic. Need for Speed uses pre-recorded sounds and then modifies them in game using pitch, reverb, etc, and they sound far better. I'm not downing TDU b/c overall it is a better game and overall experience but from a sound standpoint EA's NFS series kills Atari's TDU series (hopefully a series).
You can also compare Combat Flight Simulator 2 with Maddox's IL2 - 1946. IL2 1946 uses generated engine sound effects and while they are good they fall short. The cannons and gunfire on 1946 also are generated and are in dire need of help. CFS 2 and 3 I believe use pre-recorded sound effects. I also think that MS Flight Sim and MS Train Sim also use real world sound effects and they sound so much better.
Call of Duty 2, Call of Duty 4, MOHA, MOHA: Pacific Assault, and MOHA: Airborne also use pre-recorded sound effects. Many of these games take a week or so to go out and record actual sound effects and the end result is fantastic. Your ears, just like your eyes, are very good at picking up on repetition. Your ears are as good at picking out sounds that just don't quite sound like their real world counterparts as your eyes are at picking out textures that don't look realistic enough.
Pre-recorded sounds fare much better in games than generated ones. I believe this is also one of the reasons that musical keyboards started to support real time sampling of sounds b/c the engineers knew that regardless of how good their sound algos were...they just didn't match up to their real world counterparts. DSP sound algos are very good at generating musical instrument sounds as has been evidenced by the minds at Creative, Yamaha, etc. But when it comes to real world sounds that are a combination of many many waveforms they just don't stand up.
The human brain is a hard beast to fool when it comes to sight and sound.
From an efficiency standpoint it does not make sense to me to spend a lot of cycles running through an algorithm to produce a waveform that is readily available in the real world and sounds better to boot.
Last edited by VirtualAce; 09-25-2008 at 05:19 PM.
Direction and velocity could be scalar variables or a vector - it would definitely make sense to store the resulting X and Y deltas as a vector though (and only calculate when the angle changes).Originally Posted by mike_g
In addition, store the sines in a lookup table (which would result in some quantisation of the angle you can travel in) - the cosine would then be the entry a quarter of the table size away from the sine. Much more efficient.
Of course, that's what I would do, but I was just outlining a principle in my previous example - it's up to the coder to take it and tailor it to their application.
Current Setup: Win 10 with Code::Blocks 17.12 (GNU GCC)
